分类存储下的自动化立体仓库出入库任务调度与货位分配集成优化  

Joint Optimization for Storage and Retrieval Request Scheduling and Location Assignment in a Class-Based AS/RS

作　　者：许瑞[1,2] 妥亚方 杨鹏 XU Rui;TUO Yafang;YANG Peng(School of Business,Hohai University,Nanjing,Jiangsu 210024,China;Guangdong Provincial Key Laboratory of Brain-inspired Intelligent Computation,Southern University of Science and Technology,Shenzhen,Guangdong 518055,China;Department of Computer Science and Engineering,Department of Statistics and Data Science,Southern University of Science and Technology,Shenzhen,Guangdong 518055,China)

机构地区：[1]河海大学商学院,江苏南京210024 [2]南方科技大学广东省类脑智能计算重点实验室,广东深圳518055 [3]南方科技大学计算机科学与工程系&统计与数据科学系,广东深圳518055

出　　处：《工业工程与管理》2024年第1期159-171,共13页Industrial Engineering and Management

基　　金：国家自然科学基金面上项目(62106098,42071278,62272210);广东省类脑智能计算重点实验室(2020B121201001);深圳市自然科学基金稳定支持计划项目(20200925154942002)。

摘　　要：分类存储是自动化立体仓库(automated storage/retrieval system,AS/RS)中重要的存储策略之一。在此系统中,考虑任务的货区约束,对出入库任务调度与货位分配决策的联合优化是提升AS/RS运作效率及客户满意度的关键。本文以现实中多巷道AS/RS为研究对象,构建了以最小化任务完成时间与延迟时间加权和为目标的结合整数线性规划模型的动态规划模型,并提出了基于指派模型和文化基因算法的数学规划启发式算法。在此算法中,根据货架分区特征,引入一种基于汉明距离的新型出入库任务匹配度量指标,将入库任务排序子问题转换为指派问题,并结合文化基因算法对出库任务排序、任务分配及货位选择子问题进行优化。仿真实验验证了所提算法具有良好的优化性能,且提出的入库任务顺序优化方法可有效减少任务完成时间与延迟时间。The class-based storage assignment policy is one of the important policies in the automated storage/retrieval system(AS/RS).In this system,the joint optimization for storage and retrieval request scheduling and the location assignment plays the critical role of enhancing the operational efficiency of AS/RS and improving customer satisfaction,with the consideration of constraints on the classification of torage/retrieval requests.Using real-world multi-aisle AS/RS as the object,a dynamic programming model incorporating an integer linear programming model was established to minimize the weighted sum of request completion time and tardiness time.A matheuristic algorithm based on an assignment problem model and a memetic algorithm was designed to solve the problem.In the matheuristic algorithm,a new type of storage/retrieval request pairing metric based on Hamming distance was introduced according to the rack partition characteristics.The subproblem of storage request sequencing was transformed into an assignment problem,which integrated with the memetic algorithm to optimize the subproblem of retrieval request sequencing,allocation,and location assignment.The computational experiments show that the matheuristic algorithm has good optimization performance and the proposed optimization method of storage request sequencing can effectively reduce the completion time and tardiness time of the request.

关 键 词：AS/RS 交叉存取 分类存储 出入库任务调度 货位分配 数学规划启发式算法 

分 类 号：F715.6[经济管理—产业经济]